What is a model?

Plan for Next Week

Weeks 9 & 10

Week 9

  • Data Ethics
    • Data Context
    • Model Ethics
  • Project Work Sessions
  • Starting Your Poster

Week 10

  • Tuesday: Review Practice Final
  • Thursday: Exam 2

Finals Week

  • Poster Presentations on Saturday, March 15th from 1:10 to 2:30.
  • In our original classroom!
  • The Cal Poly print shop closes at 4pm on Friday!

:::::::::

Discussion Questions

According to O’Neil, is every mathematical model a Weapon of Math Destruction? Why or why not?

O’Neil characterizes some algorithms as “Weapons of Math Destruction.” Who is the most vulnerable to being harmed by their use?

Have you had experiences with algorithms that O’Neil’s account illuminates?

What is (or what could be) the role of data scientists in data justice?

ethics and Ethics

ethics and Ethics

ethics – a discipline concerned with what is morally good and bad and morally right and wrong


Meta-ethics (Ethics) – concerning the theoretical meaning and reference of moral propositions

Applied ethics (ethics) – concerning what a person is obligated (or permitted) to do in a specific situation or a particular domain of action

ethics

Professional Ethics

The American Statistical Association’s Ethical Guidelines for Statistical Practice are intended to help statistics practitioners make decisions ethically. Additionally, the ethical guidelines aim to promote accountability by informing those who rely on statistical analysis of the standards they should expect.

Technical Responsibilities

Ensuring analyses are valid.

  • Checking assumptions (e.g. no regression when relationship isn’t linear).
  • Matching analyses to data types (e.g., no linear regression with 0/1 response).

Reproducibility and Documentation

Being transparent about your process and its limitations.

  • Reporting insignificant results, not just significant ones.
  • Maintaining reproducibility and documenting all data processing.
  • Being clear about the scope of the conclusions.
  • Sharing your work openly when possible.

Professional Responsibilities

Performing analyses and communicating results in ways that are not misleading, even when you could get away with it.

  • No p-hacking / gathering extra data until it’s significant.
  • No switching your hypotheses after seeing the data (or running the analysis).
  • No dropping NAs unless you can justify why.

Ethics

Data Science Oath

I will respect the privacy of my data subjects, for their data are not disclosed to me that the world may know, so I will tread with care in matters of privacy and security.

I will remember that my data are not just numbers without meaning or context, but represent real people and situations, and that my work may lead to unintended societal consequences, such as inequality, poverty, and disparities due to algorithmic bias.

Note

Social Responsibilities

Considering impact, harm, and power.

  • Checking that the data collection process is not harmful (e.g. unethical experiments, uncomfortable survey questions, omitting certain groups from the study)
  • Checking that the data cleaning process is not harmful (e.g. “othering” groups)
  • Considering how conclusions will be used (e.g. who funded the study, what might they use it for, etc. - like crime prediction tools and such)

Final Project

Final Project Requirements

This document describes the requirements of the Final Project.

Your Poster will have six total sections:

  1. Introduction
  2. Methods
  3. Visualizations
  4. Model
  5. Results / Implications
  6. Ethics & Limitations

Ethics & Limitations

Model Ethics

How could your model detrimentally impact people?

Limitations

What are the limitations of your models? What are the limitations of the inferences that can be made from these data?

Poster Presentations

A Folder Full of Resources

This Google Drive folder contains:

  • poster templates
  • icons for Cal Poly (if you want to use)
  • a presentation on common poster formatting mistakes